Book a meeting
🏆 WINNER!!! - Internet Explorer
Congratulations!!!
You've won a meeting
You are visitor number 000000 and have been selected to talk to Oscar Yasunaga.
Click here to claim your prizeHot AI safety researcher in your area
Studies physics at UofT. Probes language models for truth and sycophancy. Teaches AI safety at TAISI and TARA.
📢 2026-09-15
★ Best viewed in Netscape Navigator at 800×600 ★
Every ad on this page is a real way to reach Oscar.
About Oscar
Hi, I'm Oscar. I study physics, with minors in math and statistics, at the University of Toronto. My research is in AI safety: I use interpretability tools like probes, steering and sparse autoencoders to study how language models represent concepts such as truth and sycophancy, and I look at how networks generalize through their Jacobians.
I also teach AI safety at TARA and the Toronto AI Safety Initiative. If you'd like to talk, book a coffee chat.
Blog
- Hello worldPlaceholder post. Replace or delete this file.